 Well, I hope the morning's going okay for everybody. I am really excited to bring out our next guest, which is my good friend and longtime OpenStack supporter, Brian Stevens, who's now CTO over at Google Cloud. Come on out, Brian. Hi, everybody. All right, get comfortable. Is this uncomfortable for you? Yeah, I think I can. I brought it a couple of friends. They're in the balconies as well. Yeah, so, okay, so, Brian, you were involved in OpenStack from the very early days. You helped us launch the foundation. You were on the board. You spent 13 years at Red Hat, and then you left us. What the heck, man? You went to Google? What's that all about? Well, it takes me a while to chart my path. I think about OpenStack, I missed it for a year, 18 months. It felt like a long kind of time, and then, well, because everything's about for me is like building communities and the evolution of where things are going, and so it took a little bit to see that the future was OpenStack. But then, you know, I just sort of, even at Red Hat, I spent a lot of time thinking about operational complexity, and when you look inside of Enterprise, and what does that look like? And if you ever really spend time inside of Enterprise and not just being a vendor, it's a really hard job, and they spent a lot of time doing the labors of love again and again and again, and I think what you were showing earlier around automating infrastructure is a part to alleviate that, and I just felt Public Cloud is another tool to really help alleviate and really turn things into programmable infrastructure. And so to have an opportunity to be part of that, to make that real, I had to pick up the family and move to California. Wow, so after being on the other side and telling people, it's not a big deal, just get the software and operate it, and you're like, well, I've got to actually operate it now. That's on the other side, it's not so easy, huh? No, there's some new challenges that I never quite thought about, so I'm learning every day. I mean, the part about the vendor was awesome, where you give them software, and now you actually have to run it, and it's like a goal, you only hear about your misses, right? And so the dealing with reliability, and then I think the hardest part, to be honest, is that it's constant change, right? So you're constantly changing the platform, right? Because you have to, but you've got to keep the whole ecosystem and users alive and compatible along the way, and so just sort of the dimension of sort of what that looks like has been fascinating. So one thing I'm sure at least one person in here is wondering is, does Google actually care about open source? You've been in open source most of your career, and now you're at Google, and I know they're doing stuff, but what's happening on the inside? Give us the scoop. Yeah, I think we had a saying, this is one of the things you probably shouldn't say on stage. Perfect. And I think we said, what was the word? Don't write this down, y'all. Google is the place where great open source developers go to. We said, die, but I think we heard, has never been heard from again. And we know that's not true, but it's still here. It's a pretty strong statement, but I think one of the parts of the process that I have of sort of reasoning about like whether this was a really good move for me and my family and then for the community at large, was just to really understand that Larry Page and ours were really committed to open source. And so that was the process before I accepted. And then to actually be inside of Googleplex now and actually be able to influence the open source strategy, it's really gone light years from, because the path before was publish a white paper and then boom, Hadoop's created an open source. And that's really cool, but now what you see is you're gonna see far less white papers from Google and you're just gonna see code. And I think that's, we've seen it with GRPC. Code talks. GRPC, Apache Beam. And what I was most intrigued about was, so it wasn't really around whether they were against open source, it was just sort of understanding what community engagement looks like. Like we'll look at Kubernetes. And there's a lot of opportunities. We spent a lot more time now really like, what are the areas that we should contribute? How do we create communities? And then just connecting that to strategic trust. We don't really believe a proprietary cloud wins, right? I mean, could have done that with containers. And so, there's a million reasons why for ad users, that's just bad. And so it may be the long path, but we're taking a long path and that's definitely open source. And by the way, Serra Novotny does a great job inside of Google, but out in the Kubernetes community. So shout out to her for her community management work. We can see that it's working, your efforts are working after three years. She had a long week and yes, so shout out to Serra. She's fantastic, yep. Cool, so where do you think kind of open source is heading generally? It seems like it's becoming more of the go-to model for development, but where else do you see it evolving? Well, I remember back since dating 2002 when it was called a toy in cancer and you couldn't trust it. So I think it's awesome that we're past that. I mean, everything inside of Google built on open source. And then even for us, we're seeing the model now where we have a totally different tool chain inside, which is an amazing tool chain, but you're also realizing that it doesn't kind of make sense. Like it needs to converge with what the community at large is doing. And so we're looking like, how do we really step up the contributions and build a unified community around the whole tool chain? I think one of the areas that you just, you're seeing this evolution like in what you saw in the past in proprietary as you just see, it's the constant competition. Look at like proprietary storage. It's constantly there's a new company coming in, disrupting the old company again, again and again. And open source, we're getting things behind us. It started with installers and it went to virtual machines and then it went to programmable infrastructure and then went to containers and then went to orchestration and nobody's really competing at those layers anymore. They're building on the layers. And I think the piece for me that I think that we really get to is getting from containers to services. And that has to be done in open source. Like how do you actually, like you just showed like another cool but bespoke way of deploying Cinder. But like we all should agree on like what that metadata model is for describing a service and then managing the control plane for that service. And it needs to run everywhere. It needs to run on Amazon. It needs to run on OpenStack. It needs to run anywhere. Okay, so I think what you're saying is it's one thing to compose by just kind of bringing together all these different projects and services and installing them but we don't really have a standard way of doing that. So describing the composable infrastructure in a sort of a language or a code like way that's repeatable is kind of maybe the next. Sure. That a proprietary ISV could use as well to be honest. Sure. Because the myriad of ways of doing it. And the only way you're gonna solve that is for all the disparate communities right to really get together and to rally around that OpenStack and Kubernetes and Cloud Foundry and et cetera. We'll count us in. Good. If there's some meeting, you know, I'll get all these people to come and it'll be good. Sounds good. I'll get a lot done with. Okay, so last area that I just am fascinated by I know a lot of people are following this kind of machine learning thing. The Google's code actually beat the go champion which is pretty impressive. And I think one of the things that was most exciting about that is or scary depending on how you look at the whole AI thing was that the computer actually made moves that human beings had never conceived of. Right, right. And now we see the great players learning from the computer and saying, okay, well, wonder why it made that move? Maybe we should try that. Yeah. Try that. So what do you think machine learning is going? Oh my gosh. Like the, I mean, two years ago, you didn't really hear about people talking about it so much but AI's been going on for. 50 years. Yeah, right. And so I think that's because now and I think that's because it's gone from a research effort to really solving business problems. And most of what probably a lot of people in this room do is we work on technology and infrastructure. But machine learning is the first time that I've seen that you can so concisely attach technology to solving a business problem. Because the technology's arrived. And so as an example, when you think about diabetes, it's one of the leading cause of blindness, growing causes of blindness. And so what we've done with machine learning and TensorFlow has been to actually detect that early because it's easy when you have great ophthalmologists but half a billion people around the world don't have that. And so to actually be able to use machine learning to actually be able to detect that early because it's totally curable. Like that's obvious. Like somebody's like, well, that's a problem I want to solve or cancer or fraud detection of your bank. There's some really interesting business problems that are all being solved already with machine learning. And I think that's the fascinating part is connecting that thread from problem to technology and powered by open source. Like I mean how cool is the fact that the heart of the framework, I mean it's basically the Kubernetes of machine learning is TensorFlow. The heart of the framework for challenging this, I mean really valuable. And it's happening in open source. Well, thank you for everything you've done inside of Google. I feel like we are all benefiting from Google becoming more and more open source all the time. I mean I use Android and Chrome and all these consumer technologies but with Kubernetes and TensorFlow we can see the big impact you're having internally to make Google into another great open source source of inspiration and code, not just white paper. So thank you so much for joining us Ryan. Here's, thank you. Thank you.